###############################################################################
###### Rabia Malik & Svanhildur Thorvaldsdottir
###### Are Goodwill Ambassadors Good for Business?
###### Replication of Analysis
###############################################################################


library(pastecs)
library(stargazer)
library(tidyr)
library(dplyr)
library(broom)
library(dotwhisker)
library(ggplot2)
library(ggpubr)
library(viridis)
library(stringr)


## read file generated by using the code in CleaningCode.R
## insert your own file path here
ga <- read.csv("~/GA_ReplicationData.csv")
ga_hispanic <-read.csv("~/GA_HispanicData.csv")



###############################################################################
############################   MAIN PAPER RESULTS   ###########################
###############################################################################


##########################################
##
## Table 1: Descriptive Statistics
##
#########################################

ga$amountToUNICEFTreat <- ifelse(ga$Shakira==1, ga$amountToUNICEF, NA)
ga$amountToUNICEFControl <- ifelse(ga$Shakira==0, ga$amountToUNICEF, NA)

fordesc <- data.frame(cbind(ga$amountToUNICEF,
                            ga$amountToUNICEFTreat,
                            ga$amountToUNICEFControl,
                            ga$givetoUNICEF_dum,
                            ga$givetoCause_dum,
                            ga$learnAboutUNICEF_dum,
                            ga$male,
                            ga$age,
                            ga$His_dum,
                            ga$white,
                            ga$Dem,
                            ga$Indep,
                            ga$Repub,
                            ga$college_deg,
                            ga$income_50kLess))
# for medians
stat.desc(fordesc)

# for other decriptives
stargazer(fordesc,
          covariate.labels=c("Amount Donated (to UNICEF)", 
                             "Treated group only: Amount Donated",
                             "Control group only: Amount Donated",
                             "Donate to UNICEF",
                             "Donate to Cause", 
                             "Learn about UNICEF",
                             "Male", 
                             "Age",
                             "Hispanic", 
                             "White",
                             "Democrat",
                             "Independent",
                             "Republican",
                             "College Degree",
                             "Low Income"))

# note the medians were put into the table manually in latex

############################################
##
## Figure 1: Celebrity Endorsement and Attitudes towards UNICEF
##
#############################################

# Amount donated, without controls
mod1 <- lm(amountToUNICEF ~
             Shakira, 
           data=ga)

# Amount donated, with controls
mod1b <- lm(amountToUNICEF ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

# Donate to cause, without controls
mod2 <- lm(givetoUNICEF_dum ~
             Shakira, 
           data=ga)


# Donate to cause, with controls
mod2b <- lm(givetoUNICEF_dum ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

# Donate to UNICEF, without controls
mod3 <- lm(givetoCause_dum ~
             Shakira, 
           data=ga)

# Donate to UNICEF, with controls
mod3b <- lm(givetoCause_dum ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

# Learn about UNICEF, without controls
mod4 <- lm(learnAboutUNICEF_dum ~
             Shakira, 
           data=ga)

# Learn about UNICEF, with controls
mod4b <- lm(learnAboutUNICEF_dum ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)


########################################
#### Figure 1 plots
########################################

mod1_df <-
  broom::tidy(mod1) %>% filter(term == "Shakira") %>%
  mutate(model = "DV1 Without controls",
         controls = "Without controls",
         dv = "Amount donated") %>% 
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod1b_df <-
  broom::tidy(mod1b) %>% filter(term == "Shakira") %>%
  mutate(model = "DV1 With controls",
         controls = "With controls",
         dv = "Amount donated") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod2_df <-
  broom::tidy(mod2) %>% filter(term == "Shakira") %>%
  mutate(model = "DV2 Without controls",
         controls = "Without controls",
         dv = "Donate to UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod2b_df <-
  broom::tidy(mod2b) %>% filter(term == "Shakira") %>%
  mutate(model = "DV2 With controls",
         controls = "With controls",
         dv = "Donate to UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod3_df <-
  broom::tidy(mod3) %>% filter(term == "Shakira") %>%
  mutate(model = "DV3 Without controls",
         controls = "Without controls",
         dv = "Donate to cause") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod3b_df <-
  broom::tidy(mod3b) %>% filter(term == "Shakira") %>%
  mutate(model = "DV3 With controls",
         controls = "With controls",
         dv = "Donate to cause") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod4_df <-
  broom::tidy(mod4) %>% filter(term == "Shakira") %>%
  mutate(model = "DV4 Without controls",
         controls = "Without controls",
         dv = "Learn about UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))
mod4b_df <-
  broom::tidy(mod4b) %>% filter(term == "Shakira") %>%
  mutate(model = "DV4 With controls",
         controls = "With controls",
         dv = "Learn about UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador"))

all_models <- rbind(mod1_df, mod1b_df,
                    mod2_df, mod2b_df,
                    mod3_df, mod3b_df,
                    mod4_df, mod4b_df)

p.all_base <- ggplot(all_models, aes(estimate, term, colour=controls))
p.all_base +
  geom_point(aes(shape = controls), size=2.5,
             position = position_dodge(width=0.3)) +
  geom_errorbarh(aes(xmax = estimate + 1.96*std.error,
                     xmin=estimate - 1.96*std.error,
                     height = 0),
                 position=position_dodge(width=0.3),
                 linewidth=0.75, ) +
  geom_errorbarh(aes(xmax = estimate + 1.645*std.error,
                     xmin=estimate - 1.645*std.error,
                     height = 0),
                 position=position_dodge(width=0.3),
                 linewidth=1.2) +
  geom_vline(xintercept = 0, colour="grey") +
  facet_wrap(vars(dv), nrow=1, scales="free_x") +
  scale_x_continuous(labels = scales::number_format(accuracy = 0.01)) +
  scale_y_discrete(limits = rev,
                   labels = function(x) str_wrap(x, width = 8)) + #reverses y-axis
  scale_color_viridis(discrete=TRUE, begin=0, end=0.7) +
  guides(colour = guide_legend(reverse = TRUE),
         shape= guide_legend(reverse = TRUE)) +
  theme_light() +
  theme(axis.title=element_blank(),
        legend.position = "bottom",
        legend.title= element_blank(),
        text = element_text(size = 15),
        panel.spacing.x = unit(6, "mm"),
        plot.margin = margin(0,1,0,0, "cm")) #t,r,b,l


############################################
##
## Figure 2: Celebrity Endorsement, Ethnicity, and Attitudes towards UNICEF
##
#############################################

# Amount donated, without controls
mod1_eth <- lm(amountToUNICEF ~
                 Shakira +
                 His_dum +
                 Shakira*His_dum, 
               data=ga)

# Amount donated, with controls
mod1b_eth <- lm(amountToUNICEF ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

# Donate to cause, without controls
mod2_eth <- lm(givetoUNICEF_dum ~
                 Shakira +
                 His_dum +
                 Shakira*His_dum, 
               data=ga)

# Donate to cause, with controls
mod2b_eth <- lm(givetoUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

# Donate to UNICEF, without controls
mod3_eth <- lm(givetoCause_dum ~
                 Shakira  +
                 His_dum +
                 Shakira*His_dum,  
               data=ga)

# Donate to UNICEF, with controls
mod3b_eth <- lm(givetoCause_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

# Learn about UNICEF, without controls
mod4_eth <- lm(learnAboutUNICEF_dum ~
                 Shakira  +
                 His_dum +
                 Shakira*His_dum,  
               data=ga)

# Learn about UNICEF, with controls
mod4b_eth <- lm(learnAboutUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

############################
## Figure 2 plots
############################


mod1_eth_df <-
  broom::tidy(mod1_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV1 Het Without controls",
         controls = "Without controls",
         dv = "Amount donated") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))
mod1b_eth_df <-
  broom::tidy(mod1b_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV1 Het With controls",
         controls = "With controls",
         dv = "Amount donated") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))

mod2_eth_df <-
  broom::tidy(mod2_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV2 Het Without controls",
         controls = "Without controls",
         dv = "Donate to UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))
mod2b_eth_df <-
  broom::tidy(mod2b_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV2 Het With controls",
         controls = "With controls",
         dv = "Donate to UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))

mod3_eth_df <-
  broom::tidy(mod3_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV3 Het Without controls",
         controls = "Without controls",
         dv = "Donate to cause") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))

mod3b_eth_df <-
  broom::tidy(mod3b_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV3 Het With controls",
         controls = "With controls",
         dv = "Donate to cause") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))

mod4_eth_df <-
  broom::tidy(mod4_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV4 Het Without controls",
         controls = "Without controls",
         dv = "Learn about UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))

mod4b_eth_df <-
  broom::tidy(mod4b_eth) %>%
  filter(term == "Shakira" |
           term == "His_dum" |
           term == "Shakira:His_dum") %>%
  mutate(model = "DV4 Het With controls",
         controls = "With controls",
         dv = "Learn about UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       His_dum = "Hispanic",
                       "Shakira:His_dum" = "GA x Hispanic"))

all_models_eth <- rbind(mod1_eth_df, mod1b_eth_df,
                        mod2_eth_df, mod2b_eth_df,
                        mod3_eth_df, mod3b_eth_df,
                        mod4_eth_df, mod4b_eth_df)

p.all_eth <- ggplot(all_models_eth, aes(estimate, term, colour=controls))
p.all_eth +
  geom_point(aes(shape = controls), size=2.5, position = position_dodge(width=0.3)) +
  geom_errorbarh(aes(xmax = estimate + 1.96*std.error,
                     xmin=estimate - 1.96*std.error,
                     height = 0),
                 position=position_dodge(width=0.3),
                 linewidth=0.75, ) +
  geom_errorbarh(aes(xmax = estimate + 1.645*std.error,
                     xmin=estimate - 1.645*std.error,
                     height = 0),
                 position=position_dodge(width=0.3),
                 linewidth=1.2) +
  geom_vline(xintercept = 0, colour="grey") +
  facet_wrap(vars(dv), nrow=1, scales="free_x") +
  scale_y_discrete(limits = rev) + #reverses y-axis
  scale_color_viridis(discrete=TRUE, begin=0, end=0.7) +
  guides(colour = guide_legend(reverse = TRUE),
         shape= guide_legend(reverse = TRUE)) +
  theme_light() +
  theme(axis.title=element_blank(),
        legend.position = "bottom",
        legend.title= element_blank(),
        text = element_text(size = 15))


############################################
##
## Figure 3: Celebrity Endorsement, Gender, and Attitudes towards UNICEF
##
#############################################

# Amount donated, without controls
mod1_gen <- lm(amountToUNICEF ~
                 Shakira +
                 male +
                 Shakira*male, 
               data=ga)


# Amount donated, with controls
mod1b_gen <- lm(amountToUNICEF ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*male,
                data=ga)


# Donate to cause, without controls
mod2_gen <- lm(givetoUNICEF_dum ~
                 Shakira +
                 male +
                 Shakira*male, 
               data=ga)


# Donate to cause, with controls
mod2b_gen <- lm(givetoUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  male*Shakira,
                data=ga)


# Donate to UNICEF, without controls
mod3_gen <- lm(givetoCause_dum ~
                 Shakira +
                 male +
                 male*Shakira, 
               data=ga)


# Donate to UNICEF, with controls
mod3b_gen <- lm(givetoCause_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  male*Shakira,
                data=ga)


# Learn about UNICEF, without controls
mod4_gen <- lm(learnAboutUNICEF_dum ~
                 Shakira +
                 male +
                 male*Shakira, 
               data=ga)


# Learn about UNICEF, with controls
mod4b_gen <- lm(learnAboutUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  male*Shakira,
                data=ga)



############################
## Figure 3 plots
############################

mod1_gen_df <-
  broom::tidy(mod1_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV1 Gen Without controls",
         controls = "Without controls",
         dv = "Amount donated") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))
mod1b_gen_df <-
  broom::tidy(mod1b_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV1 Gen With controls",
         controls = "With controls",
         dv = "Amount donated") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))

mod2_gen_df <-
  broom::tidy(mod2_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV2 Gen Without controls",
         controls = "Without controls",
         dv = "Donate to UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))

mod2b_gen_df <-
  broom::tidy(mod2b_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV2 Gen With controls",
         controls = "With controls",
         dv = "Donate to UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))

mod3_gen_df <-
  broom::tidy(mod3_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV3 Gen Without controls",
         controls = "Without controls",
         dv = "Donate to cause") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))

mod3b_gen_df <-
  broom::tidy(mod3b_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV3 Gen With controls",
         controls = "With controls",
         dv = "Donate to cause") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))

mod4_gen_df <-
  broom::tidy(mod4_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV4 Gen Without controls",
         controls = "Without controls",
         dv = "Learn about UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))

mod4b_gen_df <-
  broom::tidy(mod4b_gen) %>%
  filter(term == "Shakira" |
           term == "male" |
           term == "Shakira:male") %>%
  mutate(model = "DV4 Gen With controls",
         controls = "With controls",
         dv = "Learn about UNICEF") %>%
  relabel_predictors(c(Shakira = "Goodwill Ambassador",
                       male = "Male",
                       "Shakira:male" = "GA x Male"))


all_models_gen <- rbind(mod1_gen_df, mod1b_gen_df,
                        mod2_gen_df, mod2b_gen_df,
                        mod3_gen_df, mod3b_gen_df,
                        mod4_gen_df, mod4b_gen_df)


p.all_gen <- ggplot(all_models_gen, aes(estimate, term, colour=controls))
p.all_gen +
  geom_point(aes(shape = controls), size=2.5, position = position_dodge(width=0.3)) +
  geom_errorbarh(aes(xmax = estimate + 1.96*std.error,
                     xmin=estimate - 1.96*std.error,
                     height = 0),
                 position=position_dodge(width=0.3),
                 linewidth=0.75, ) +
  geom_errorbarh(aes(xmax = estimate + 1.645*std.error,
                     xmin=estimate - 1.645*std.error,
                     height = 0),
                 position=position_dodge(width=0.3),
                 linewidth=1.2) +
  geom_vline(xintercept = 0, colour="grey") +
  facet_wrap(vars(dv), nrow=1, scales="free_x") +
  scale_y_discrete(limits = rev) + #reverses y-axis
  scale_color_viridis(discrete=TRUE, begin=0, end=0.7) +
  guides(colour = guide_legend(reverse = TRUE),
         shape= guide_legend(reverse = TRUE)) +
  theme_light() +
  theme(axis.title=element_blank(),
        legend.position = "bottom",
        legend.title= element_blank(),
        text = element_text(size = 15))


###############################################################################
############################   APPENDIX   #####################################
###############################################################################

###########################
##
## Table A2: Number of respondents by treatment
##
###########################

# Note this table was made by hand, using the following calculations

sum(ga$Shakira_GE==1)
sum(ga$Shakira_GE==1 & ga$female==1)
sum(ga$Shakira_GE==1 & ga$female==0)

sum(ga$Expert_GE==1)
sum(ga$Expert_GE==1 & ga$female==1)
sum(ga$Expert_GE==1 & ga$female==0)

sum(ga$Shakira_DC==1)
sum(ga$Shakira_DC==1 & ga$female==1)
sum(ga$Shakira_DC==1 & ga$female==0)

sum(ga$Expert_DC==1)
sum(ga$Expert_DC==1 & ga$female==1)
sum(ga$Expert_DC==1 & ga$female==0)

###########################
##
## Table A3: ANOVA tests for balance
##
###########################

# The variable Treat_4 has all four treatment groups to run anova analyses on

# Male
reg_male <- lm(male ~ Treat_4,
               data=ga)
anova(reg_male) 

#Age
reg_age <- lm(age ~ Treat_4,
              data=ga)
anova(reg_age) 

#Hispanic
reg_hispanic <- lm(His_dum ~ Treat_4,
                   data=ga)
anova(reg_hispanic)

#White
reg_white <- lm(white ~ Treat_4,
                data=ga)
anova(reg_white) 

#Democrat
reg_dem <- lm(Dem ~ Treat_4,
              data=ga)
anova(reg_dem) 

#Independent
reg_ind <- lm(Indep ~ Treat_4,
              data=ga)
anova(reg_ind) 

#Republican
reg_rep <- lm(Repub ~ Treat_4,
              data=ga)
anova(reg_rep) 

#College Degree
reg_college <- lm(college_deg ~ Treat_4,
                  data=ga)
anova(reg_college)

#Low Income
reg_inc <- lm(income_50kLess ~ Treat_4,
              data=ga)
anova(reg_inc) 

#Have children
reg_child <- lm(child ~ Treat_4,
                data=ga)
anova(reg_child) 

#Freq. of news reading
reg_news <- lm(news_check_num ~ Treat_4,
               data=ga)
anova(reg_news) 

#Country Knowledge
reg_countryKnow <- lm(SC1 ~ Treat_4,
                      data=ga)
anova(reg_countryKnow) 

#Organization Knowledge
reg_OrgKnow <- lm(SC3 ~ Treat_4,
                  data=ga)
anova(reg_OrgKnow)

###########################
##
## Table A4: Dependent variable descriptives by treatment group
##
###########################

## Shakira_Girls'Education:

fordesc.data <- ga[ga$Shakira_GE==1,]
fordesc <- data.frame(cbind(fordesc.data$amountToUNICEF,
                            fordesc.data$givetoUNICEF_dum,
                            fordesc.data$givetoCause_dum,
                            fordesc.data$learnAboutUNICEF_dum
                            ))
# for medians
stat.desc(fordesc)

# for other decriptives
stargazer(fordesc,
          covariate.labels=c("Amount Donated", 
                             "Donate to UNICEF",
                             "Donate to Cause", 
                             "Learn about UNICEF"
                             ))

# note the medians were put into the table manually in latex

## Shakira_Displaced Children:

fordesc.data <- ga[ga$Shakira_DC==1,]
fordesc <- data.frame(cbind(fordesc.data$amountToUNICEF,
                            fordesc.data$givetoUNICEF_dum,
                            fordesc.data$givetoCause_dum,
                            fordesc.data$learnAboutUNICEF_dum
))
# for medians
stat.desc(fordesc)

# for other decriptives
stargazer(fordesc,
          covariate.labels=c("Amount Donated", 
                             "Donate to UNICEF",
                             "Donate to Cause", 
                             "Learn about UNICEF"
          ))

# note the medians were put into the table manually in latex

## Expert_Girls' Education:

fordesc.data <- ga[ga$Expert_GE==1,]
fordesc <- data.frame(cbind(fordesc.data$amountToUNICEF,
                            fordesc.data$givetoUNICEF_dum,
                            fordesc.data$givetoCause_dum,
                            fordesc.data$learnAboutUNICEF_dum
))
# for medians
stat.desc(fordesc)

# for other decriptives
stargazer(fordesc,
          covariate.labels=c("Amount Donated", 
                             "Donate to UNICEF",
                             "Donate to Cause", 
                             "Learn about UNICEF"
          ))

# note the medians were put into the table manually in latex

## Expert_Displaced Children:

fordesc.data <- ga[ga$Expert_DC==1,]
fordesc <- data.frame(cbind(fordesc.data$amountToUNICEF,
                            fordesc.data$givetoUNICEF_dum,
                            fordesc.data$givetoCause_dum,
                            fordesc.data$learnAboutUNICEF_dum
))
# for medians
stat.desc(fordesc)

# for other decriptives
stargazer(fordesc,
          covariate.labels=c("Amount Donated", 
                             "Donate to UNICEF",
                             "Donate to Cause", 
                             "Learn about UNICEF"
          ))

# note the medians were put into the table manually in latex

## note that table was manually assembled in latex based on these values


###########################
##
## Table A5: Celebrity Endorsement and Attitudes towards UNICEF: Full Results
##
###########################

mod1 <- lm(amountToUNICEF ~
             Shakira, 
           data=ga)

mod1b <- lm(amountToUNICEF ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

mod2 <- lm(givetoUNICEF_dum ~
             Shakira, 
           data=ga)

mod2b <- lm(givetoUNICEF_dum ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

mod3 <- lm(givetoCause_dum ~
             Shakira, 
           data=ga)

mod3b <- lm(givetoCause_dum ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

mod4 <- lm(learnAboutUNICEF_dum ~
             Shakira, 
           data=ga)

mod4b <- lm(learnAboutUNICEF_dum ~
              Shakira +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

# table with all four DVs 
stargazer(mod1, mod1b, mod2, mod2b, mod3, mod3b, mod4, mod4b, type="latex",
          title="Celebrity Endorsement and Attitudes towards UNICEF: Full Results",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Goodwill Amb.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A6: Celebrity Endorsement, Ethnicity, and Attitudes towards UNICEF: Full Results
##
###########################
m1_e <- lm(amountToUNICEF ~
                 Shakira +
                 His_dum +
                 Shakira*His_dum, 
               data=ga)

m1b_e <- lm(amountToUNICEF ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

m2_e <- lm(givetoUNICEF_dum ~
                 Shakira +
                 His_dum +
                 Shakira*His_dum, 
               data=ga)

m2b_e <- lm(givetoUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

m3_e <- lm(givetoCause_dum ~
                 Shakira  +
                 His_dum +
                 Shakira*His_dum,  
               data=ga)

m3b_e <- lm(givetoCause_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

m4_e <- lm(learnAboutUNICEF_dum ~
                 Shakira  +
                 His_dum +
                 Shakira*His_dum,  
               data=ga)

m4b_e <- lm(learnAboutUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*His_dum,
                data=ga)

stargazer(m1_e, m1b_e, m2_e, m2b_e,m3_e, m3b_e, m4_e, m4b_e,  
          type="latex",
          title="Celebrity Endorsement, Ethnicity and Attitudes towards UNICEF",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Goodwill Amb.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income",
                             "GA$\times$Hisp."),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)

###########################
##
## Table A7: Celebrity Endorsement, Ethnicity re-weighted, and Attitudes towards UNICEF
##
###########################

m1_er <- lm(amountToUNICEF ~ Shakira +
                    His_dum +
                    Shakira*His_dum,
                  data=ga_hispanic)

m1b_er <- lm(amountToUNICEF ~ Shakira +
                     His_dum +
                     Shakira*His_dum +
                     male +
                     age +
                     white +
                     Dem +
                     Indep +
                     college_deg +
                     income_50kLess, 
                   data=ga_hispanic)

m2_er <- lm(givetoUNICEF_dum ~
                    Shakira +
                    His_dum +
                    Shakira*His_dum, 
                  data=ga_hispanic)

m2b_er <- lm(givetoUNICEF_dum ~
                     Shakira +
                     male +
                     age +
                     His_dum +
                     white +
                     Dem +
                     Indep + 
                     college_deg +
                     income_50kLess +
                     Shakira*His_dum,
                   data=ga_hispanic)

m3_er <- lm(givetoCause_dum ~
                    Shakira  +
                    His_dum +
                    Shakira*His_dum,  
                  data=ga_hispanic)

m3b_er <- lm(givetoCause_dum ~
                     Shakira +
                     male +
                     age +
                     His_dum +
                     white +
                     Dem +
                     Indep + 
                     college_deg +
                     income_50kLess +
                     Shakira*His_dum,
                   data=ga_hispanic)

m4_er <- lm(learnAboutUNICEF_dum ~
                    Shakira  +
                    His_dum +
                    Shakira*His_dum,  
                  data=ga_hispanic)

m4b_er <- lm(learnAboutUNICEF_dum ~
                     Shakira +
                     male +
                     age +
                     His_dum +
                     white +
                     Dem +
                     Indep + 
                     college_deg +
                     income_50kLess +
                     Shakira*His_dum,
                   data=ga_hispanic)

# table generation
stargazer(m1_er, m1b_er, m2_er, m2b_er, m3_er, m3b_er, m4_er, m4b_er, type="latex",
          title="Celebrity Endorsement, Ethnicity reweighted, and Attidues Towards to UNICEF",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Shakira",
                             "Hispanic",
                             "Male",
                             "Age",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Deg.",
                             "Low Income",
                             "Shak.*Hisp."),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)

###########################
##
## Table A8: Celebrity Endorsement, Gender, and Attitudes towards UNICEF
##
###########################

m1_g <- lm(amountToUNICEF ~
                 Shakira +
                 male +
                 Shakira*male, 
               data=ga)

m1b_g <- lm(amountToUNICEF ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  Shakira*male,
                data=ga)

m2_g <- lm(givetoUNICEF_dum ~
                 Shakira +
                 male +
                 Shakira*male, 
               data=ga)

m2b_g <- lm(givetoUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  male*Shakira,
                data=ga)

m3_g <- lm(givetoCause_dum ~
                 Shakira +
                 male +
                 male*Shakira, 
               data=ga)

m3b_g <- lm(givetoCause_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  male*Shakira,
                data=ga)

m4_g <- lm(learnAboutUNICEF_dum ~
                 Shakira +
                 male +
                 male*Shakira, 
               data=ga)

m4b_g <- lm(learnAboutUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess +
                  male*Shakira,
                data=ga)

# table with all four DVs 
stargazer(m1_g, m1b_g, m2_g, m2b_g, m3_g, m3b_g, m4_g, m4b_g, 
          type="latex",
          title="Celebrity Endorsement, Gender, and Attitudes towards UNICEF",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Goodwill Amb.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income",
                             "Goodwill Amb.*Male"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A9: Celebrity Endorsement, Issues, and Attitudes towards UNICEF
##
###########################

m1_i <- lm(amountToUNICEF ~
             Shakira_GE +
             Shakira_DC +
             Expert_DC, 
           data=ga)

m1b_i <- lm(amountToUNICEF ~
              Shakira_GE +
              Shakira_DC +
              Expert_DC +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

m2_i <- lm(givetoUNICEF_dum ~
             Shakira_GE +
             Shakira_DC +
             Expert_DC, 
           data=ga)

m2b_i <- lm(givetoUNICEF_dum ~
              Shakira_GE +
              Shakira_DC +
              Expert_DC +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

m3_i <- lm(givetoCause_dum ~
             Shakira_GE +
             Shakira_DC +
             Expert_DC, 
           data=ga)

m3b_i <- lm(givetoCause_dum ~
              Shakira_GE +
              Shakira_DC +
              Expert_DC +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

m4_i <- lm(learnAboutUNICEF_dum ~
             Shakira_GE +
             Shakira_DC +
             Expert_DC, 
           data=ga)

m4b_i <- lm(learnAboutUNICEF_dum ~
              Shakira_GE +
              Shakira_DC +
              Expert_DC +
              male +
              age +
              His_dum +
              white +
              Dem +
              Indep + 
              college_deg +
              income_50kLess,
            data=ga)

stargazer(m1_i, m1b_i, m2_i, m2b_i, m3_i, m3b_i, m4_i, m4b_i, 
          type="latex",
          title="Celebrity Endorsement, Issues, and Attitudes towards UNICEF",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("G.Amb GE",
                             "G.Amb. DC",
                             "Expert DC",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)

###########################
##
## Table A10: Attention Robustness Check: Celebrity Endorsement and Attitudes towards UNICEF
##
###########################

# subsetting OUT those respondents who SAW the celebrity
#   but did not accurately recall her
ga_correct <- subset(ga, incorrect==0) 

m1_in <- lm(amountToUNICEF ~
                   Shakira,
                 data=ga_correct)

m1b_in <- lm(amountToUNICEF ~
                    Shakira +
                    male +
                    age +
                    His_dum +
                    white +
                    Dem +
                    Indep + 
                    college_deg +
                    income_50kLess,
                  data=ga_correct)

m2_in <- lm(givetoUNICEF_dum ~
                   Shakira, 
                 data=ga_correct)

m2b_in <- lm(givetoUNICEF_dum ~
                    Shakira +
                    male +
                    age +
                    His_dum +
                    white +
                    Dem +
                    Indep + 
                    college_deg +
                    income_50kLess,
                  data=ga_correct)

m3_in <- lm(givetoCause_dum ~
                   Shakira, 
                 data=ga_correct)

m3b_in <- lm(givetoCause_dum ~
                    Shakira +
                    male +
                    age +
                    His_dum +
                    white +
                    Dem +
                    Indep + 
                    college_deg +
                    income_50kLess,
                  data=ga_correct)

m4_in <- lm(learnAboutUNICEF_dum ~
                   Shakira, 
                 data=ga_correct)

m4b_in <- lm(learnAboutUNICEF_dum ~
                    Shakira +
                    male +
                    age +
                    His_dum +
                    white +
                    Dem +
                    Indep + 
                    college_deg +
                    income_50kLess,
                  data=ga_correct)

# table with all four DVs 
stargazer(m1_in, m1b_in, m2_in, m2b_in, m3_in, m3b_in, m4_in, m4b_in, 
          type="latex",
          title="Attention Robustness Check: Celebrity Endorsement and Attitudes towards UNICEF",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Goodwill Amb.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A11: Likeability Robustness Check: Celebrity Endorsement and Attitudes towards UNICEF
##
###########################

# subset OUT those who either do not know Shakira or dislike her
ga_like_know <- subset(ga, DK_Shakira==0 & Dislike_Shak==0)

m1_lk <- lm(amountToUNICEF ~
                      Shakira,
                    data=ga_like_know)

m1b_lk <- lm(amountToUNICEF ~
                       Shakira +
                       male +
                       age +
                       His_dum +
                       white +
                       Dem +
                       Indep + 
                       college_deg +
                       income_50kLess,
                     data=ga_like_know)

m2_lk <- lm(givetoUNICEF_dum ~
                      Shakira, 
                    data=ga_like_know)

m2b_lk <- lm(givetoUNICEF_dum ~
                       Shakira +
                       male +
                       age +
                       His_dum +
                       white +
                       Dem +
                       Indep + 
                       college_deg +
                       income_50kLess,
                     data=ga_like_know)

m3_lk <- lm(givetoCause_dum ~
                      Shakira, 
                    data=ga_like_know)

m3b_lk <- lm(givetoCause_dum ~
                       Shakira +
                       male +
                       age +
                       His_dum +
                       white +
                       Dem +
                       Indep + 
                       college_deg +
                       income_50kLess,
                     data=ga_like_know)

m4_lk <- lm(learnAboutUNICEF_dum ~
                      Shakira, 
                    data=ga_like_know)

m4b_lk <- lm(learnAboutUNICEF_dum ~
                       Shakira +
                       male +
                       age +
                       His_dum +
                       white +
                       Dem +
                       Indep + 
                       college_deg +
                       income_50kLess,
                     data=ga_like_know)

# table with all four DVs 
stargazer(m1_lk, m1b_lk, m2_lk, m2b_lk, m3_lk, m3b_lk, m4_lk, m4b_lk, 
          type="latex",
          title="Likeability Robustness Check: Celebrity Endorsement and Attitudes towards UNICEF",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Goodwill Amb.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A12: Celebrity Endorsement and Attitudes towards UNICEF - Robustness Check
##
###########################

# exclude "never donors" from UNICEF donation and Cause donation

ga_charitable <- subset(ga, neverDonate_Cause==0)
ga_charitable2 <- subset(ga, neverDonate_UNICEF==0)

# amount to UNICEF at all - robustness (dummy)
m2_nd <- lm(givetoUNICEF_dum ~
                 Shakira, 
               data=ga_charitable2)

m2b_nd <- lm(givetoUNICEF_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess,
                data=ga_charitable2)

m3_nd <- lm(givetoCause_dum ~
                 Shakira, 
               data=ga_charitable)

m3b_nd <- lm(givetoCause_dum ~
                  Shakira +
                  male +
                  age +
                  His_dum +
                  white +
                  Dem +
                  Indep + 
                  college_deg +
                  income_50kLess,
                data=ga_charitable)

stargazer(m2_nd, m2b_nd, m3_nd, m3b_nd, type="latex",
          title="Celebrity Endorsement and Attitudes towards UNICEF - Robustness Check",
          style="ajps",
          summary=TRUE,
          column.labels=c("Donate to UNICEF", "Donate to Cause"),
          column.separate=c(2, 2),
          covariate.labels=c("Goodwill Amb.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A13: Political Leaning and Issue Interaction Effect (Across Treatments)
##
###########################

# subsetting to those who clearly identify as Dem or Rep.
strong_pol <- subset(ga, Dem==1|Repub==1)

m1_sp<- lm(amountToUNICEF ~
                Dem +
                GirlsEduc +
                Dem*GirlsEduc,
              data=strong_pol)

m1b_sp<- lm(amountToUNICEF ~
                 Dem + 
                 GirlsEduc + 
                 Dem*GirlsEduc +
                 male +
                 age +
                 His_dum + 
                 white +
                 college_deg +
                 income_50kLess,
               data=strong_pol)

m2_sp<- lm(givetoUNICEF_dum ~
                Dem +
                GirlsEduc +
                Dem*GirlsEduc,
              data=strong_pol)

m2b_sp<- lm(givetoUNICEF_dum ~
                 Dem + 
                 GirlsEduc + 
                 Dem*GirlsEduc +
                 male +
                 age +
                 His_dum + 
                 white +
                 college_deg +
                 income_50kLess,
               data=strong_pol)

m3_sp<- lm(givetoCause_dum ~
                Dem +
                GirlsEduc +
                Dem*GirlsEduc,
              data=strong_pol)

m3b_sp<- lm(givetoCause_dum ~
                 Dem + 
                 GirlsEduc + 
                 Dem*GirlsEduc +
                 male +
                 age +
                 His_dum + 
                 white +
                 college_deg +
                 income_50kLess,
               data=strong_pol)

m4_sp<- lm(learnAboutUNICEF_dum ~
                Dem +
                GirlsEduc +
                Dem*GirlsEduc,
              data=strong_pol)

m4b_sp<- lm(learnAboutUNICEF_dum ~
                 Dem + 
                 GirlsEduc + 
                 Dem*GirlsEduc +
                 male +
                 age +
                 His_dum + 
                 white +
                 college_deg +
                 income_50kLess,
               data=strong_pol)

# table with all four DVs 
stargazer(m1_sp, m1b_sp, m2_sp, m2b_sp, m3_sp, m3b_sp, m4_sp, m4b_sp, 
          type="latex",
          title="Political Leaning and Issue Interaction Effect (Across treatments)",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Democrat",
                             "Girls' Educ.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "College Degree",
                             "Low Income",
                             "Dem.*GE"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A14: Political Leaning-Issue Interaction Effect within Celebrity Treatment
##
###########################

# Further subsetting to those who identify as Dem or Repub AND were in the celebrity treatment
strong_pol2 <- subset(strong_pol, Shakira==1)

m1_pi<- lm(amountToUNICEF ~
                     Dem +
                     GirlsEduc +
                     Dem*GirlsEduc,
                   data=strong_pol2)

m1b_pi<- lm(amountToUNICEF ~
                      Dem + 
                      GirlsEduc + 
                      Dem*GirlsEduc +
                      male +
                      age +
                      His_dum + 
                      white +
                      college_deg +
                      income_50kLess,
                    data=strong_pol2)

m2_pi<- lm(givetoUNICEF_dum ~
                     Dem +
                     GirlsEduc +
                     Dem*GirlsEduc,
                   data=strong_pol2)

m2b_pi<- lm(givetoUNICEF_dum ~
                      Dem + 
                      GirlsEduc + 
                      Dem*GirlsEduc +
                      male +
                      age +
                      His_dum + 
                      white +
                      college_deg +
                      income_50kLess,
                    data=strong_pol2)

m3_pi<- lm(givetoCause_dum ~
                     Dem +
                     GirlsEduc +
                     Dem*GirlsEduc,
                   data=strong_pol2)

m3b_pi<- lm(givetoCause_dum ~
                      Dem + 
                      GirlsEduc + 
                      Dem*GirlsEduc +
                      male +
                      age +
                      His_dum + 
                      white +
                      college_deg +
                      income_50kLess,
                    data=strong_pol2)

m4_pi<- lm(learnAboutUNICEF_dum ~
                     Dem +
                     GirlsEduc +
                     Dem*GirlsEduc,
                   data=strong_pol2)

m4b_pi<- lm(learnAboutUNICEF_dum ~
                      Dem + 
                      GirlsEduc + 
                      Dem*GirlsEduc +
                      male +
                      age +
                      His_dum + 
                      white +
                      college_deg +
                      income_50kLess,
                    data=strong_pol2)

# table with all four DVs 
stargazer(m1_pi, m1b_pi, m2_pi, m2b_pi, m3_pi, m3b_pi, m4_pi, m4b_pi, 
          type="latex",
          title="Political Leaning Interaction Effect within Celebrity Treatment",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated", "Donate to UNICEF", "Donate to Cause", "Learn about UNICEF"),
          column.separate=c(2, 2,2,2),
          covariate.labels=c("Democrat",
                             "Girls' Educ.",
                             "Male",
                             "Age",
                             "Hispanic",
                             "White",
                             "College Degree",
                             "Low Income",
                             "Dem.*GE"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###########################
##
## Table A15: Donations to UNICEF: Real v Hypothetical Donation Question
##
###########################
mod1_real <- lm(amountToUNICEF ~
                  Shakira + 
                  Real_dum +
                  Shakira*Real_dum,
                data=ga)

mod1b_real <- lm(amountToUNICEF ~
                   Shakira +
                   Real_dum +
                   male +
                   age +
                   white +
                   Dem +
                   Indep +
                   college_deg +
                   income_50kLess +
                   Shakira*Real_dum,
                 data=ga)

stargazer(mod1_real, mod1b_real, type="latex",
          title="Donations to UNICEF: Real v Hypothetical Donation Question",
          style="ajps",
          summary=TRUE,
          column.labels=c("Amount Donated"),
          column.separate=c(2),
          covariate.labels=c("Shakira",
                             "Real",
                             "Male",
                             "Age",
                             "White",
                             "Democrat",
                             "Independent",
                             "College Degree",
                             "Low Income",
                             "Shakira*Real"),
          dep.var.labels.include=FALSE,
          digits=2,
          align=TRUE,
          keep.stat=c("n", "adj.rsq")
)


###############################################################################
############################   END   ##########################################
###############################################################################